A menu bar icon appears whenever its HTTP server is running, giving you quick access to server controls, settings, and the URL you need to connect. And because some databases should simply never touch an AI, the Exclude from Chat & MCP option lets you exclude entire databases at once — enabled automatically for encrypted and revision-proof databases.

Beyond the MCP server, DEVONthink 4.3 Herschel brings sweeping AI improvements. The chat assistant now redacts more sensitive information, e.g., phone numbers, credit card numbers, authentication tokens, and labeled secrets, before anything is sent to a language model, a protection that extends to AI smart actions and summarize and transform functions. It supports the Apple Intelligence Foundation model for simple tasks, and you can choose which AI model handles automatic tagging. We have also updated numerous AI models, including Claude Opus 4.7, GPT 5.5, Gemini 3.5 Flash, Mistral Small 4 and Medium 3.5, Flux 2, and many OpenRouter models.

Replacing MultiMarkdown, our new Markdown parser and renderer supports callouts, citations, captioned tables, and improved CriticMarkup. It is also more compatible with AI-generated output. Markdown documents know about smart quotes and dashes, and converting them to plain text strips all formatting and renders MathJax in readable form.

Finally, new desktop widgets for macOS Sequoia and later let you quickly view or access your Reading List or Favorites, load a workspace, or read the useful tips we regularly publish on our blog. For Server users, the web interface gets a dialog for opening provided item links and lets you reveal, move, duplicate, and replicate items from the context menu. The See Also inspector has a new section for classifying documents, and we have rebuilt the PDF viewer with a table of contents, page thumbnails, and full-text search across all browsers.

💬 Discuss this post in our user forum.

If you use custom metadata in DEVONthink Pro and Server or DEVONthink To Go, you’ve probably used text-based attributes like Single-Line Text. These attributes are defined in DEVONthink’s Settings > Data and commonly accessed in Tools > Inspectors > Info: Data. You can use each type in the ways you want, but here are the features and limitations of each:

• Single-Line Text remembers and lets you choose values you’ve previously typed. You can also add default values in the Data settings. That clears the memory of those previously entered values and only items you’ve defined will be offered. However, unlike a Set, you still can add ad-hoc information too. Note, these impromptu entries will not be remembered so if you want to use the value again, add it to the attribute’s settings.
• Multi-Line Text is exactly as described: text over multiple lines but without any formatting. To create new paragraphs, press ⌥⏎.
• The Rich Text type is similar to the multi-line option, but does allow formatting, e.g., bold and italic text. Create new paragraphs by simply pressing ⏎. You can even press ⌘T to open the Font panel and switch fonts.
• The Identifier is the simplest and strictest type: you enter text, with no formatting, without multiple lines, and no memory of previously entered values. This is often used for static, single-purpose information like software license keys, policy numbers, and similar static values.
• With a Set, you define the only values you can choose in the popup. This lets you strictly control the choices you (or others) have. If you need a new option to choose, add it to the attribute’s settings.

In DEVONthink To Go, you’ll open Settings > Data for definitions. To open a document’s Data inspector, open the Info popup, then tap ℹ︎ next to General, and choose Data.

💬 Discuss this post in our user forum.

If you have ever wondered where your tags are or didn’t know each of your DEVONthink databases has its own Inbox, or maybe felt like you’re getting lost scrolling to find groups, you should know about unified elements. Unified items gather some items of the same type in the Navigate sidebar, the leftmost pane on a main window. This is similar to how Apple Mail, for instance, has Inboxes in its Favorites sections. Toggle unified elements in DEVONthink’s Settings > General > Interface > Sidebar.

Unify Inboxes and Tags

Every database has its own Inbox and Tags group. If they aren’t unified, these appear alongside the other groups in the database. Enable Unify inboxes and all inboxes, including the Global Inbox, are collected in the Globals > Inboxes section. Enable Unify tags and all Tags groups appear in Globals > Tags.

When Inboxes and/or Tags are unified, clicking on those top-level items in the Globals section displays a union of the contents of every Inbox or Tags group it contains. Click the disclosure triangle to only access items of individual, open databases.

Unify Databases

When databases are unified, they are listed in the Open Databases section and can all be expanded there — the databases and their groups are all unified in one sidebar section.

Disabling this option does two things: (1) Open databases’ names are shown in Open Databases giving you quick access to the root of each database, and (2) a section for the selected database appears in the sidebar, listing only the groups in that database, isolating its groups.

Unify Elements in DEVONthink To Go

For mobile users, in DEVONthink To Go you can unify Inboxes or Tags via the View Options button (with three descending lines) on the databases screen or in Settings > Interface & Chat.

💬 Discuss this post in our user forum.

Callouts are highlighted blocks of text in a document showing notes or asides. For example, in a textbook or an online tutorial, you have likely come across highlighted blocks pointing out tips, notes, or important information. In DEVONthink To Go 4.1 Lagrange, our new Markdown renderer now supports easily creating them in your Markdown documents.

In Markdown, callouts are a type of blockquote but use a special syntax: > [!type-of-callout] Optional Title. Make sure there is no space between the exclamation point and the callout type. In the next line, you start with another > and then type your text. Here are the supported callouts (you can see what they look like in the screenshot above):

Important

> [!important] Boating Tip
> Remember this alliterative phrase: Red right, returning…

Note

This is the same as [!info].

> [!note] 
> The minimum system requirement is macOS 13.5 Ventura.

Warning

You can create multi-line callouts by placing a blockquote character on each line.

> [!warning] Beware of coral snakes
> When you see a snake, repeat this rhyme:  
> Red touch black, friend of Jack  
> Red touch yellow, kill a fellow.

Caution

> [!caution]
> Slow down for yellow lights.

Tip

> [!tip] Accounting Trick 01
> Errors in numerical transposition are always divisible by 9.

Currently, this feature is only available in DEVONthink To Go, but we’re also bringing it to DEVONthink in a future release.

💬 Discuss this post in our user forum.

While you’re reading through a PDF or other document in DEVONthink, you may run into words or phrases and wonder what other documents may contain them. You could start a new search, then copy and paste the text, or re-type it. But there is a more efficient method.

Select a word or phrase in a document. Control-click it and choose Search »…« or Search Selected Phrase, depending on whether you selected just one word or multiple words. DEVONthink starts a search in the same main window.

• If you don’t find anything interesting in the search results, clear the search field by clicking the × at the right end, and you will return to your document.
• If you do find a result you want to look at, double-click it to open it in a new document window. In the new window, the search terms will be highlighted. If the Search inspector is open, it will immediately search the document and show the occurrences. If not, press ⌘F to open the inspector with the search term ready to use.

If you use the Pro or Server edition, you have a second option available. Double-click a word in the Tools > Inspectors > Concordance inspector, and it will initiate a search in your databases.

💬 Discuss this post in our user forum.

iOS widgets are used by many people for reading little bits of at-a-glance information, viewable outside an application. Introduced with the latest version, DEVONthink To Go 4.1 Lagrange now also offers a small suite of widgets to add to your home screen.

The widgets provide a view into a few commonly used features and help keep you updated with your news. The widgets also let you quickly access your most important data right from your home screen. To add widgets, long press the background on your home screen until the Edit button appears in the upper-left corner. Tap it and choose Add Widget. Select DEVONthink and scroll through the choices:

• Unread Articles lists the total number of articles in RSS or Atom feeds you haven’t had a chance to read yet. As your feeds may be spread across multiple locations, tapping this widget only opens DEVONthink To Go.
• Recent Feeds lists the name and total articles of the three feeds with the most unread content.
• Inbox Items displays the total number of documents in an inbox. Tap the widget to jump right into the inbox.
• Favorites lists the four most recently accessed items — groups or documents — from your Favorites global smart group.
• Recently Viewed lists your four most recently viewed documents.
• The Reading List has two widgets. The smaller On the Reading List widget shows the total number of documents in the Reading List. The larger one shows the totals but also the top three items in the list.

For widgets listing items, tap to open them quickly in your database.

Now you can see how much you’ve missed in the news, jump to that document you were working on, or get back to clearing out your Reading List!

💬 Discuss this post in our user forum.

Our new Markdown renderer replaces Multi­Markdown 6 and delivers more reliable previews and broader syntax support. It combines the most popular features of Multi­Markdown, Common­Mark, GitHub Flavored Markdown, and more.

Version 4.1 also supports BibTeX documents, merges multiple images, including TIFFs, into a single PDF document, and let’s you open text documents and PDFs directly in editing mode, saving you an extra tap every time you want to make changes. We’ve revisited the group selector, too, which now sports a search field and keyboard navigation. If you’re relying on the geolocation saved with each document you can now open a geolocation in Apple and Google Maps or edit it if it’s incorrect.

Of course, we’ve also improved the overall reliability and performance, and fixed bugs and glitches.

You can find a complete list of all improvements in Settings > Help > Read Me inside the app. The New & Improved page also highlights what’s new in each feature release. We recommend this update to all users of DEVONthink To Go. The App Store will either install the update automatically for you or you can find it in the Updates area of the App Store app. If you are still using an older generation of DEVONthink To Go, version 4 is a new app and you’ll need to install it manually. DEVONthink To Go 4 is included with every license for DEVONthink 4 Pro or Server for Mac as long as it receives updates.

💬 Discuss this post in our user forum.

If you’ve ever needed to consolidate content from multiple documents into one, you likely have used copy and paste. But if you make edits, they have to be made to all involved documents. Wouldn’t it be nice if you could keep the documents separated but show them as a single document? You can with a Markdown feature in DEVONthink: file transclusion.

File transclusion displays the content of other documents as if they’re part of the Markdown document you’re working on. Typically used for other Markdown documents, it also supports plain text, rich text, PDF documents, and sheets. You can trans­clude images, though we recommend using the regular Markdown syntax instead. It is even possible to trans­clude audio and video documents. The document you want to trans­clude just needs to be in the same database as the Markdown document.

To add a transclusion, use one of these methods:

• Start typing double curly braces {{ and the name of the document you want to trans­clude. An autocompletion popup appears, so you can select a document in the same database. Select one and press Return to insert the name and close the braces.
• Type the name and extension and braces manually. This also works in DEVONthink To Go.
• Hold the Control key while dragging and dropping into the source to insert a transclusion.

The contents of the trans­cluded documents immediately appear in the preview. Change the order in the source and it changes the preview.

But note, the trans­cluded content is only a visual representation. You cannot select and edit it. To make changes to it, you need to open the document. These changes will be will be immediately reflected in the preview. Also be aware file transclusion isn’t supported in all third-party Markdown applications.

💬 Discuss this post in our user forum.

It’s not uncommon for us to get a question about finding documents with or without tags in their DEVONthink or DEVONthink To Go databases. People often approach the problem with a Tags property, but that isn’t optimal. Here’s how to correctly search for them.

If you click into the toolbar search in DEVONthink and press ⏎, the advanced search options appear. In the first popup in the second line of the search, select Item. In the field next to it (the operator popup), select is or isn’t. And in the third popup, select Tagged. You have just constructed a proper search for tagged or untagged items. Also, you can see the raw query in the toolbar search: item:tagged or item:!tagged, which you can always type in manually.

You can also select other options like duplicated, replicated, etc. In DEVONthink for Mac, the operator popup also includes contains and does not contain with some other attributes, e.g., Aliases, URL, Thumbnail, etc. DEVONthink To Go doesn’t support the contains / does not contain operators at this time, but it is on our list of improvements.

To do the same kind of search in DEVONthink To Go, tap the search icon and enter the search terms manually or tap the Action button (⋯) and choose Advanced Search to use the predicate editor.

💬 Discuss this post in our user forum.

Joe Kissell from Take Control Books has updated his ebook Take Control of DEVONthink 4. It is now up to date with DEVONthink 4.2.2 Cassini and DEVONthink To Go 4.0.6 Kepler. Joe explains, for example, how to export reminders from DEVONthink to Apple Reminders or how to use the new AI search settings. He also updated many screenshots to match DEVONthink’s new Liquid Glass interface for macOS Tahoe.

Download Take Control of DEVONthink 4 for free from our Handbooks and Extras page or directly from Take Control Books.

💬 Discuss this post in our user forum.

In DEVONthink, you can encrypt databases for particularly private or sensitive documents. When you open your encrypted database, you will be prompted to enter your encryption key. This macOS dialog has a small checkbox, which may be enabled by default: Remember password in keychain. So the next time you want to open the database, DEVONthink will automatically use the stored key from your keychain and won’t ask you to enter it. While this may be convenient, you may not want it for privacy reasons. Do this to remove your key from the keychain:

1. Close the database.
2. In the Finder, do a Spotlight search for Keychain and open the Keychain Access application. Do not open the Passwords application when prompted. You want to use Keychain Access. On modern versions of macOS, you will likely need to enter your administrative password as well.
3. Select the login section on the left.
4. Click the search field in the upper right and search for spars. This will show you stored encryption keys for disk images.
5. If you find the key for your database, select it, choose Edit > Delete, and confirm.
6. Quit Keychain Access.
7. Reopen your encrypted database and you should again be prompted for the encryption key to open it. Make sure you disable Remember password in keychain to avoid running into the same issue.

This also applies to DEVONthink’s revision-proof databases.

💬 Discuss this post in our user forum.

Recently, we had a question about the Today global smart group in DEVONthink and why it’s not in DEVONthink To Go. To keep things simple and functional, we chose to include a smaller set of very commonly used smart groups by default. But that’s no problem at all, because you can easily create your own smart groups.

Smart groups show items that match criteria you specify, e.g., PDF documents added in the past week. They are essentially saved custom searches, so this is the standard way to create them:

1. Tap the search field or magnifying glass icon to start a search.
2. Type in your search terms, choose the scope of the search, e.g., All or Group, and inspect the results.
3. Tap the Action button (…) and choose Save.
   • If your scope was All, this will prompt you to save a global smart group. This appears on the databases screen.
   • If your scope was any other option, you will save a local smart group. This appears in the current database.

There is an alternative method for creating a local smart group: go into the desired database, tap the Add button (+), and choose Smart Group. Give it a name, add your criteria, and save it. And yes, local smart groups sync with DEVONthink on the Mac.

So, to create a Today global smart group, run a search with the criteria added:today kind:document, save it, and name it Today.

Note: On the databases screen on the iPhone, you can’t choose a scope. You’re not in a specific database, so you will automatically create a global smart group.

💬 Discuss this post in our user forum.

AI inside DEVONthink can be a powerful tool. But when it comes to AI, some users have security concerns, such as possible prompt injections. So, what exactly are they, and are they a risk in DEVONthink?

A prompt injection happens when text hidden inside a document is crafted to circumvent AI with its own commands, like one that says, “Ignore the previous prompt and do this instead…” The danger isn’t limited to triggering unwanted commands — injections can also silently distort AI responses such as summaries or evaluations. In addition, injections can occur not only in your own documents but also in web search results and downloaded pages, which is why searches and downloads should be restricted and only enabled when needed.

This isn’t a DEVONthink problem specifically but an issue for any document processing AI system. This is why AI models are increasingly designed to be more resistant to these attacks, and the best precaution is to be mindful about what you add to your database. While in a well-curated personal database the practical risk is low, there are features in DEVONthink to limit or thwart these injections:

• AI responses are automatically stripped of unexpected or potentially dangerous HTML content (such as scripts or frames) that could result from prompt injections.
• Even if you allow downloads, the Chat assistant, Chat smart actions, and AppleScript commands automatically disable downloading web pages while processing a document. The Chat assistant also disables downloads after accessing the content, annotations, or comments of a document.

You can also limit what the AI is allowed to access, both inside and outside your databases. However, keep in mind that stricter limits may reduce the AI’s ability to assist you:

• Enabling Search > Sources: Web in the settings or Search: Web in the Chat’s Options is often unnecessary and increases exposure to prompt injections via search results or downloaded pages. Enable it only as needed.
• If you do enable web searching, consider also disabling the AI > Chat > Assistant: Allow download of web pages and Load remote images automatically options. Disabling remote images also prevents prompt injections from silently transferring data to remote servers.
• You could disable AI > Search: Database and also AI > Chat > Assistant: Allow property & content changes. You can temporarily enable database searching in the Chat assistant as needed.
• When using the Chat assistant or Chat-related smart actions, e.g., Chat - Query, set the Documents popup to Without Document, unless you are specifically dealing with documents.

We don’t intend to cause undue concern, but to help you make better-informed choices in how you use AI.

💬 Discuss this post in our user forum.

It is not uncommon for us to get support inquiries saying they installed DEVONthink or DEVONthink To Go but it’s not syncing. DEVONthink does not automatically sync by default because we believe that you should decide for yourself whether and what you want to sync. Software shouldn’t make such decisions for you. Here are things to consider.

Before setting up the sync, think about…

• whether you want to sync, as sync is always optional.
• what database(s) to sync. Not every database needs to (or should) be synchronized.
• which devices to sync to, as not every database needs to be on every device. This is especially true on mobile.
• which sync method to use, noting each device must be using the same sync method/location to sync the same database.
• whether you want to use encryption when syncing. You must make this choice before you sync the first time. If you do use encryption, you must use the same encryption key on other devices syncing with the same sync method.

Sync is discussed in the built-in Help and user handbook for both applications. There is a Sync Explained section in the In & Out chapter for DEVONthink and the Sync chapter for DEVONthink To Go. Once you decide what sync method you’ll use, each method has its own section with a description, technical considerations, and simple step-by-step instructions for setting up a new or an existing sync location. If you only have one of our applications, just follow the steps to create a new location on one device, then follow the steps for an existing location on the other device(s). Setting up sync in DEVONthink and DEVONthink To Go is pretty similar but if you want to be sure you can check the corresponding chapters in both handbooks.

You can find many more blog posts on sync-related topics here on our blog. The free Take Control of DEVONthink also includes a dedicated chapter, Sync DEVONthink Databases, and there are several entries about synchronization in our FAQ.

If you do run into an issue, you can consult the Troubleshooting > Sync Issues section of the Help or start a support ticket, if you really get stuck.

💬 Discuss this post in our user forum.

If you want to take notes about a document in your DEVONthink database, there is a feature that does exactly that: the annotation. This is a specialized note document linked to another document. It is most often used for taking long-form notes about PDFs, but can be used with any document in your database. We already introduced them in a blog post some time ago. However, here is a short recap of how to create them, as well as some additions to the previous post.

Select a document in your database. To create an annotation, select Data > Annotate & Remind and choose one of the default templates. Alternatively, open the Tools > Inspectors > Info: Annotations & Reminders inspector, click the Annotations button and choose from the New from Template menu. Or just start typing in the annotation field and the document is created for you.

Above the annotation field are buttons to open the annotation in its own window (the arrow), create a backlink to the document you’re writing about (the chain), or insert a quote from selected text (the quote). In the Annotations menu, you will find similar commands but also some new AI-related ones:

• Insert Summary via Chat: The selected chat model summarizes the current document or its selected text. Click insert to add in the summary into the annotation.
• Insert Transcription: Inserts recognized text from an image or speech from a media document, as explained in this blog post about AI speech transcription.

You can find more information on all annotation commands in the Inspectors > Info pane > Annotations & Reminders > Annotation Files section in the built-in Help or user handbook.

Lastly, you can configure some default settings for the annotations in Settings > Files > General:

• Choose whether to create annotation documents in the same group as your document or in a shared Annotations group at the root of this database.
• Choose whether annotations should “follow” the originating document if it is moved to another database.
• Choose whether the annotation document’s name should be updated if you rename the originating document. If the annotation has a backlink to the original document, you must also enable the WikiLinks > Update name of item links automatically setting to change the link’s text.
• The annotation document can be either Markdown or rich text. Choose one of these options or let DEVONthink decide based on the originating document’s type.

💬 Discuss this post in our user forum.

We have blogged about all the sidebars in a DEVONthink window — the Navigate sidebar, the Reading List, and the Import sidebar of the Pro and Server edition — except one: the Extras sidebar. While it may seem like just a simple pane in the interface, we believe it’s a useful place to get some additional information from us.

This last of the sidebars appears as an ellipsis (•••). It functions as a newsroom and a library of DEVONthink-related information. If you see a notification about a new blog post, etc., it’s directing you to this sidebar. Connected to online resources, like our blog, it provides up-to-date information without you having to leave the application.

In the Support section, the Tips & Tricks subsection has a large number of current instructional blog posts, e.g., our tips, like this one, to read at your leisure. You can also click the article’s title to open it in your browser. The Scripts, Templates, and Smart Rules subsections mirror the installable extras from the Support Assistant window. Each item includes a brief description and a live hyperlink so you can install extras directly from this window.

Lastly, the DEVONtechnologies section lists our promotions or those we’ve participated in, as well as application updates we’ve announced.

So if you are curious what the latest information or tip is, switch to the Extras and see what’s new.

💬 Discuss this post in our user forum.

Many people are interested in running local AI models but hardware limits what models you can run. Smaller models are less reliable, larger models are too much for most machines. But the popular AI application Ollama now offers access to some large models online, ones that can be used in DEVONthink or DEVONthink To Go.

Ollama often is an entry point for people new to local AI. Recently, they started offering cloud models, providing access to many models including large LLMs like Mistral-large-4:675b (675 billion parameters!). In our testing, some of these models are very performant, keeping pace with commercial ones. To set up Ollama and these models in the higher editions of DEVONthink, do the following:

1. Create an account on Ollama.com.
2. Click on Keys then Add API Key. Enter a name to distinguish the key, e.g., Ollama Remote then press Generate API Key. Copy this key and store it somewhere safe as you will not be able to copy it again.
3. If you haven’t downloaded the Ollama application, do so now and launch it.
4. Control-click Ollama’s menubar icon, open the Settings, and sign into your Ollama account. This will register the device so you know what devices can access your Ollama cloud models. While not required, you can also copy and store the public key as a reference.
5. Go into DEVONthink’s Settings > AI and choose Ollama (Remote) in the Provider popup.
6. Copy and paste your API key into the API Key field.
7. In the API URL field, enter: https://ollama.com/api/chat. You should immediately see the Model popup display a model. Click on it to choose the online model you want to use.

To use this in DEVONthink To Go, follow steps 5-7 in the app’s Settings > AI.

If you’re interested, Ollama also offers paid services to access more models, etc. Bear in mind, AI models vary in capabilities, so one may be able to “read a document” and generate a Markdown document with a response, but not all of them can. We don’t control what they can and can’t do. You’ll need to test and become familiar with the model(s) you want to use.

💬 Discuss this post in our user forum.

DEVONthink user Stephen Millard has created an automation for importing documents. On his blog ThoughtAsylum, he explains how he uses AppleScript folder actions and the app Dropzone to import and file documents in certain databases and groups in DEVONthink.

His post is a good example of an automation that Stephen created through tinkering and trial and error. It is interesting to see how he describes his work process, where he tried different approaches and finally found a good solution. Stephen shares the script in his post so you can use it or study it in detail if you’re interested in scripting. Check out Stephen’s blogpost here.

💬 Discuss this post in our user forum.

This month four years ago the Russian Federation started the invasion of Ukraine and brought war back to Europe. Since then, soldiers and civilians alike have died day after day. Now, the cold winter and the targeted destruction of civil energy infrastructure is making life in Ukraine almost unbearable. Pipes burst from frozen water, people who are unable to leave, die in their icy homes.

We are just a small team of developers working from the US, the EU, and the UK, and as a company we don’t take a political stance. But, we cannot ignore a humanitarian catastrophe happening in our neighborhood.

We have today donated 20 % of our projected February revenue to two humanitarian organizations sending teams to Ukraine repairing damaged homes and infrastructure, and providing covers, food, and generators.

We often talk about document links, e.g., in the Documents > Document Linking section of DEVONthink's Help. The See Also inspector exposes potentially related documents but there are other relationships between items not as easily seen. In version 4, these connections are visualized via its Graph view.

The Graph view is available in the Pro and Server editions of DEVONthink and driven by its internal AI and other metadata. Depending on the properties of the selected document, the connections the graph visualizes could include: tags, See Also suggestions, mentions, item links, or WikiLinks. For example, if you have a document containing WikiLinks to other documents, you can see and show or hide those connections.

You can access the Graph View via Tools > Inspectors > AI: Graph or via the detachable Tools > Graph popover. This is a context-aware feature so as you change documents, it updates the connections and options. Double-click nodes to view the connected item. In the inspector, this also changes the central node of the graph. Zoom in or out with ⌘+ or ⌘-. You can even use the arrow keys and the Return key to navigate the nodes and switch to selected items in the graph. You’ll find other commands in the context menu when you Control-click a node.

Be aware, the graph is necessarily limited in it’s scope. It is not intended to display branches beyond the immediate connections to the central node. Also, the display is static at this time so you can’t drag the nodes to different positions.

PS: This is a built-in feature, separate from the scripted version in version 3 we blogged about some years ago.

💬 Discuss this post in our user forum.